Why the buy-side is winning the battle for the hottest machine learning talent
There's a new and increasingly important job in financial services - the head of machine learning. Hedge funds, investment banks and large asset managers are creating these roles and battling for expertise in big data, data science and AI, but the buy-side is currently winning the battle.
Man Group's appointment of William Ferreira in the newly-created role of head of machine learning follows J.P. Morgan's recruitment of Geoffrey Zweig from Microsoft in February. It's indicative of a broader trend towards hiring in AI experts across the financial sector, both in (very) senior roles and lower down the ladder. While banks like Morgan Stanley have been hiring quants and systematic trading professionals, it's hedge funds like Squarepoint Capital and Balyasny Asset Management which have been recruiting more in this area, according to recruiters.
"Hiring managers are looking for excellent academics and people who’ve worked in a tier-one or two shop,” said Andrew Cronin, vice president and the head of systematic trading at GQR. “A quantitative skill set is interesting to firms on the buy side and the sell side.
While investment banks have long used big data analytics to streamline risk management, operations and cost-cutting, more are looking at data science, big data analytics and AI in a revenue-generation capacity, much as quant hedge funds have been using machine learning for algorithmic execution.
“Banks have realized that machine learning algorithms can do various tasks faster than any human ever could, such as managing the data that the group holds,” he said. “On the buy side, they want people who can develop trading strategies based on machine learning algorithms that can find patterns using structured data sets, mainly quant traders, PMs and people who’ve worked with algorithmic execution products."
Buy-side and sell-side hiring managers are looking for a strong combination of quantitative skills. In addition to computer science, they look for candidates who studied disciplines such as applied math, stats, electrical engineering, various sciences, including physics, operations research and game theory. A major in economics or finance is not technical enough for these types of roles.
“They want someone who can really drill into one area of the business and optimize that, someone who can solve mathematical brain teasers, but also has a very strong programming background in C++, Java, Python, R, Matlab, Arthur Whitney’s kdb or really any object-oriented programming language,” Cronin said.
The buy-side is set to be transformed by the move towards machine learning. This currently involves buying in huge quantities of third-party data, such as credit card receipts or information on crop yields, and attempting to extract information that could give them an edge over the competition. There are a couple of problems, however - the data is still, largely, incredibly messy and there are only a handful of people on the market who really crunch the data to generate trading opportunities.
“They want to hire people who can extract and analyze structured and unstructured data,” said Cronin. “Machine learning algorithms make decisions in a far more applied and systematic way than a human brain ever could – the processing power of a computer much more powerful than a human being – so being able to build algorithms is more important than any other traits.”
Hedge funds and asset managers account for the bulk of recruitment, says Robin Isaacson, the owner of recruiters Isaacson Search Co. (ISC), and are looking in more obscure places for talent. Some hiring managers are asking for candidates who have successfully competed in Kaggle programming contests, which require knowledge of data science and machine learning, she says.
“A lot of hedge fund firms have lost money in discretionary trading, so they want to change to systematic and compete – they’re always trying to find new ways of making money,” she said. “They’re all looking for people with a computer science or statistical background and coders with programming experience in either C++ or Python."
Photo credit: PeopleImages/GettyImages